Page 7 - Fiorentino_alii_2008

VARIABILITY IN LAND SNAILS 815

Figure 5. A, Mantel correlogram (distance class: 25 km) assessed by Euclidean distances from the Z-matrix. B, Mantel
correlogram (distance class: 25 km) assessed by Euclidean distances from mitochondrial DNA sequences distance matrix
(maximum likelihood estimated).

Plots of the -ln likelihood scores over generation grouping the south-eastern populations plus Favig-
time showed that stable parameter estimates were nana specimens; clade 6, grouping the large ‘M.
obtained after approximately 200 trees (= 20 000 gen- muralis’ group but, in this case, there was no clear
erations). Therefore, only trees sampled after this relationships with geographic sampling. Interestingly,
burn-in period were used to determine posterior prob- Marettimo Island was resolved as a deeply divergent
abilities of model parameters, branch lengths and and basal lineage.
clades, by generating a consensus tree with PAUP* Mantel correlograms based on population scores for
(50% majority-rule). mtDNA sequences (Fig. 5B) showed a quite different
Parsimony analysis from the combined dataset pro- spatial pattern than did morphology: from 0 to
duced 48 equally parsimonious trees of 548 steps 225 km, a steep decline in rz from significant positive
(CI = 0.604, RI = 0.635) differing in minor branch values to negative ones and, from 50 to 200 km,
arrangement within the ‘M. muralis’ populations. BI fluctuations in rz around zero but, again, with a steep
and MP recovered basically the same relationships of decrease in rz to significant negative values by the
lineages and clades (Fig. 6). 225 km distance class. Examination of all pairs of
All reconstructions failed to recognize the five populations involved in this class showed that they
species as well shell shape groups as monophyletic were the extreme south-eastern populations, as for
lineages and sequences were mainly resolved in dis- shell shape analysis.
tinct geographic lineages, but with some exceptions.
Five separate unresolved lineages were recognized.
Within them, at least six well supported geographic DISCUSSION
assemblages were identified: clade 1, grouping
central-western populations; clade 2, grouping north- GENETIC VERSUS MORPHOLOGICAL DIFFERENTIATION
western extremity population; clade 3, grouping The phylogenetic relationships in Sicilian Marmorana
northern chain (Madonie Mountains) populations; inferred from molecular data did not match morpho-
clade 4, grouping south-western populations; clade 5, logical (shell and genitalia) patterns. Unexpectedly,

© 2008 The Linnean Society of London, Biological Journal of the Linnean Society, 2008, 94, 809–823

Downloaded from https://academic.oup.com/biolinnean/article-abstract/94/4/809/2701256
by guest
on 18 November 2017